A Search for Spectroscopic Binaries Among Herbig Ae/Be Stars?,??

ASTRONOMY & ASTROPHYSICS MAY I 1999, PAGE 429 SUPPLEMENT SERIES Astron. Astrophys. Suppl. Ser. 136, 429–444 (1999)

A search for spectroscopic binaries among Herbig Ae/Be stars?,??

P. Corporon and A.-M. Lagrange Laboratoire d’Astrophysique de l’Observatoire de Grenoble, BP. 53, F–38 041 Grenoble Cedex, France

Received March 4; accepted November 17, 1998

Abstract. We present the results of a spectroscopic sur- binaries (see a detailed review in Clarke 1996), but obser- vey of binaries among 42 bright (mV < 11) Herbig vations are needed to constrain further these models. Ae/Be stars in both hemispheres. Radial velocity vari- Secondly, a fundamental role of binary studies is to ations were found in 7 targets, 4 are new spectroscopic allow the direct determination of physical parameters. binaries. The Li i 6 708 A˚ absorption line (absent feature Noticeably, the stellar mass is only accessible through in simple HAeBe stars spectra) indicates the presence of the observation of gravitationally bound pairs of stars, by a cooler companion in 6 HAeBe spectrum binaries, 4 of straight application of gravitational law. which are new detections. Few stars classified as possible Main-Sequence (MS) binary stars are overall quite well Herbig Ae/Be stars are not confirmed as such. studied. However, as orbits of binary systems evolve with While for short-period (P<100 days) spectroscopic time, it is mandatory to derive the properties of the sys- binaries, the observed binary frequency is 10%, the true tems during the pre-Main Sequence (PMS) phase. A major spectroscopic binary frequency for Herbig Ae/Be stars issue is to quantify the binary frequency fb (the probabil- maybeashighas35%. ity that a given object is multiple, Reipurth & Zinnecker 1993) for young multiple objects, their separation distri- Key words: binaries: spectroscopic — stars: early-type — bution as well as their mass ratio. stars: pre-Main Sequence — stars: statistics Recent studies have shown that more than half of T Tauri stars, young stars having a mass M<1.5M , are members of a binary or multiple system (Mathieu 1992; Leinert et al. 1993; Reipurth & Zinnecker 1993; Prosser et al. 1994; Simon et al. 1995; Ghez et al. 1993, 1. Introduction 1997; Brandner et al. 1996; Padgett et al. 1997). Whether there is an overabundance of low-mass PMS binaries ver- 1.1. Multiplicity and evolution sus MS binaries is still a matter of debate, due to the difficulties to compare both statistics obtained with differ- Multiplicity is a major issue in stellar astrophysics. Firstly, ent approaches. Not only the employed techniques are dif- binary stars are very common among Main Sequence (MS) ferent, (MS stars were mainly spectroscopically searched stars: half of the MS field stars are known to belong to for, while PMS binaries were searched with high angular multiple systems (see Garmany et al 1980 for O type resolution imaging), but also the wavelength domains are stars; Abt et al. 1990 for B; Nordström et al. 1997 for different (optical observations predominate for MS stars, F; Duquennoy & Mayor 1991 for G; Mayor et al. 1992 for while infrared (IR) surveys of the young objects, mostly K; Leinert et al. 1997 for M). Thus, any stellar forma- embedded in dark regions, were used ipso facto). The only tion theory must explain this large abundance of multiple systematic effort aimed at finding new low-mass pre-Main systems. Various mechanisms have been proposed to form Sequence binaries with visible spectroscopy dates back from Mathieu (1992). Send offprint requests to:P.Corporon ? Based on observations collected at the European Southern The question is then whether the stellar density may Observatory (ESO), La Silla, Chile and at the Observatoire de have an effect on the formation of multiple systems. MS Haute–Provence (OHP), Saint–Michel l’Observatoire, France. stars are thought to have been formed in OB associa- ?? Table 1 only available in electronic form at CDS via tions or in dense clusters (Miller & Scalo 1978; Lada anonymous ftp to cdsarc.u-strasbg.fr (130.79.128.5) or via et al. 1991), while PMS stars in T associations are formed http://cdsweb.u-strasbg.fr/Abstract.html in lower density environments which might enhance the 430 P. Corporon and A.-M. Lagrange: A search for HAeBe spectroscopic binaries production of binaries. However, Brandner et al. (1996) 1.3. Aim of the present paper observed T Tauri stars in OB and T associations and concluded that the binary frequency is the same for In this paper, we report the first results of a systematic MS and PMS low mass stars in the range of separation high resolution spectroscopic search for HAeBe binaries. 120 − 1 800 AU (except for the Taurus–Auriga star form- It is part of an extensive survey made on Herbig Ae/Be ing region). Moreover, Padgett et al. (1997) using HST stars, the other facet being high angular resolution imag- observations of PMS stars in dense clusters recently found ing of IR companions of HAeBe stars using Adaptative a comparable binary frequency for dense clusters and low- Optics systems (Bouvier et al. 1998; Corporon 1998). The density star-forming regions. They thus claim that multi- two approaches are complementary: while the former gives plicity is not influenced by the local stellar density (at least access to the study of short orbital periods (P = few hours in regions with densities ranging between 40 and 5 000 to few months) of double stars, the latter covers the do- stars pc−3). main of longer periods (P = many years). Since the pioneering work of Mathieu (1992), selec- The present paper is structured as follows: in Sect. 2, tive mass determinations for some low-mass PMS binaries we describe our sample, the instruments used as well as have been obtained (Padgett & Stapelfeldt 1994; Welty our observing strategy. We present in Sect. 3 the spectra 1995; Figueiredo 1997), but the sample of young multiple of some known and new spectroscopic HAeBe binaries. systems with orbital and physical parameters determined Notes on individual sources are given in Sect. 4 and in must be enlarged in order to test the early stages of stellar Sect. 5 we discuss our results. evolutionary models. 2. Observations of the HAeBe binaries 1.2. Multiplicity of intermediate-mass PMS stars 2.1. The sample As pointed out by Hillebrand (1994), the intermediate- mass PMS stars, namely Herbig Ae/Be (HAeBe) stars The observed Herbig Ae/Be stars were extracted from are found in various environments: in dense star form- Table 1 of Thé et al. (1994) catalogue. We obtained high ing regions (containing few tens of HAeBe plus a myriad resolution spectra for 42 objects with mV < 11, consist- of T Tauri stars), in lower density groups of young stel- ing of our principal HAeBe sample. This sample represents lar objects (2 to 5 HAeBe stars sharing the same birth about 70% of the HAeBe candidates with mV < 11 from place) and in isolated molecular cores (where a central Table 1 of Thé et al. (1994) catalogue. Most of the remain- HAeBe star and embedded young lower mass stars are ing stars were not observed either because they were al- found). Studying the binary frequency among HAeBe ob- ready well studied or because they are strong photometric jects may help to better understand the relation between variables and out of our limit of our observing capabilities the direct environment and the multiplicity status of the at their minimum brightness. stars during their earlier formation stages. A sub-group of other HAeBe candidates listed in Besides reinforcing the binary frequency estimate, Tables 2 to 5 from the catalogue of Thé et al. (1994) were HAeBe binaries study is of great interest because direct also studied (sample T2–T5). They were included in our mass determination are fervently required to test the survey because they could be observed in parallel to our stellar evolution models for young intermediate-mass main program. Those stars were observed not only to test stars. the presence of a companion, but also to precise, when To date, however, the binarity status of Ae/Be Herbig possible, their spectral type or to test their belonging to stars has been far less surveyed, probably because these the Herbig Ae/Be stellar class. stars form a class less homogeneous than T Tauri stars We address in this paragraph the possible bias by the (Thé et al. 1994). Few recent studies using infrared imag- so-called Branch effect. Branch (1976) pointed out that a ing (Li et al. 1994; Leinert et al. 1997b; Pirzkal et al. 1997) magnitude-limited sample favors the detection of double- have found a binary frequency (although based on limited lined spectroscopic binary (SB2). In other words, some samples) in excess by a factor 2 versus A/B type MS stars, binary stars could have been selected because their total (by considering G type MS stars degree of multiplicity fb, magnitude is below the magnitude limit, whereas indi- spectroscopically determined by Duquennoy and Mayor vidual stars are fainter than the mV limit. Such systems 1991 identical through the Main Sequence, as explained should be removed when establishing the binary frequency by Leinert et al. 1993). of the sample. However, in our case with high mass pri- Up to now, the pilot study made on the eclipsing and maries, unless the mass ratio is close to unity, the sec- spectroscopic triple system TY CrA (Lagrange et al. 1993; ondary component of a spectroscopic binary system con- Corporon et al. 1994, 1996; Beust et al. 1997, see also tribute to few percents of the combined flux. We consider Casey et al. 1993, 1995, 1998) is the only work that led the example of the double-lined spectroscopic binary sys- to the first direct determination of masses for a HAeBe tem TY CrA, adopting the most recent physical param- multiple system. eters determination by Casey et al. (1998): the primary P. Corporon and A.-M. Lagrange: A search for HAeBe spectroscopic binaries 431

(late-B type) has an effective temperature TI of about the spectrograph. The final wavelength calibration accu- −1 12 000 K and a radius RI =1.8R , while the secondary racy is 2 km s . The spectra were then transformed into (late-G type) has TII ≈ 4 900 K and RII =2.1R .The the heliocentric rest frame and normalized to unity. Bad flux ratio between the two stars is pixels or cosmic rays were removed by-hand. All the data ESO MIDAS 4 2 reduction steps were performed with the FII TII × RII software. = 4 2 =3.8%. FI TI × RI ELODIE´ , located at the 1.93 m OHP telescope, Thus, in a HAeBe binary system with a primary of spec- is a fiber-fed échelle spectrograph. The detector is a tral type A or B and a lower-mass companion, the pri- 1 024 × 1 024 Tektronics CCD. 67 orders are simultane- ously recorded, giving in a single exposure a spectrum mary will be responsible for most of the observed flux. ˚ Therefore, the Branch effect is not significant in such between 3 906 and 6 811 A, at a resolution of 42 000. The cases. When both members of the spectroscopic HAeBe optical layout as well as the reduction procedure can be found in Baranne et al. (1996). The fiber diameter binary system have similar masses, and hence comparable 00 luminosity, special attention should be paid to see whether is 2 on the sky and thorium and science spectra were or not one member of the system would have been in our obtained separately. Sky background was estimated in sample or not. the inter-order space. Typical exposure times range from 0.5 to 1 hour. The final velocity calibration is better than Noticeably, we are faced with the large disparity in the −1 location of the HAeBe type stars (some of them have a 1kms . Noticeably, a program at the telescope auto- poorly determined distance). A distance limited criterion matically performs the data reduction: no further work as considered by Duquennoy & Mayor (1991) is hardly but heliocentric velocity correction and normalization to conceivable here to select a sufficiently large sample and unity is required to have a set of homogeneous spectra. to determine a reliable binary frequency for Herbig Ae/Be stars. 2.2.2. Southern hemisphere

2.2. Observing journal Southern HAeBe stars of our sample were observed with ´ The spectroscopic survey, initiated in 1994, was carried the CES (Coudé Echelle spectrograph) fed by the 1.4 m out in the two hemispheres, using three different instru- CAT telescope (La Silla Observatory, Chile). Most of ments which characteristics and data reduction procedures the observations were made under remote control from are described hereafter. the ESO headquarters in Garching bei München. From December 1994 to October 1995, the short camera config- uration and the CCD #9 was used. Then we used the long ˚ 2.2.1. Northern hemisphere camera and the new CCD #38 allowing a larger (≈ 60 A) spectral coverage in the Li i 6 708 A˚ region. The resolv- Spectra of northern hemisphere HAeBe stars were taken ing power was set to 60 000. Typical integration time was with ELODIE´ and AURELIE´ , two spectrographs of the 1 − 1.5 hour, the rectangular slit dimensions were ranging Observatoire de Haute–Provence (OHP), south of France. from 5 − 1000 × 1 − 200. The data reduction procedure was ´ AURELIE´ is a high resolution spectrograph mounted identical to the one followed for the AURELIE data (ex- at the Coudé focus of the 1.52 m telescope. A detailed cept that the 2D- spectra were averaged perpendicularly description is given in Gillet et al. (1994). The detector to the dispersion). at the time of the observations was a Thomson double The time distribution of the observations is a combina- array and two different gratings were usually employed: tion of high-frequency coverage (less than few days) and #7 with a resolution R ≈ 38 000 at λ = 6 500 Aand#5˚ low-frequency coverage (greater than 1 year) over the final (2nd order) with R ≈ 70 000 at λ = 6 500 A.˚ Typical ex- 3 years of the survey. Table 1 (only available in electronic posure time was 1 − 1.5 hour for our target stars, with form at the CDS) presents the observed stars. a circular entrance hole of 300 on the sky. A continuous Note: given the slit/fiber sizes and the extreme see- light-source provided the flat field exposures to correct ing values during the various observations (0.7 − 200), we the instrumental response. The flat-field images were cho- estimate that a binary separated by less than 1.500 has sen to have a level similar to that of the science exposures been observed while integrating the flux from both com- and were repeated each night. After subtraction of the ponents: such a system is thus considered “as a single bias and dark current (measured routinely during each target” in our spectroscopic observations. On the other night), the science exposure was divided by an average hand, the primary HAeBe (brightest component in V )of of ten suited flat-field images. Thorium and argon lamps a visual binary with separation greater than 1.500 was in- were used for wavelength calibration, and numerous ex- dependently observed (i.e. without integrating the flux of posures were taken each night to monitor the stability of its companion), when seeing conditions allowed it. 432 P. Corporon and A.-M. Lagrange: A search for HAeBe spectroscopic binaries

2.3. Spectral analysis Table 2. HAeBe candidates detected as spectroscopic binaries

Star Li i line Vrad separation vsini We present here the two methods used to spectroscopically −1 detection variations ρvis (km s ) identify a HAeBe binary star. HK Ori Y N 0.3500 a,b 150 V380 Ori Y N 0.1500 a,b 200 V586 Ori Y N 1.0000 a 160 e i ˚ 2.3.1. Search for Li 6 708 A absorption NX Pup Y N 0.1300 a,d 120 e HD 203024 Y N 0.3000 a 125 ˚ Martin (1994) quantitatively showed that the Li i 6 707.8 A MWC 863 Y N 1.1000 a,c 100 f resonance doublet can be used to detect T Tauri com- TY CrA Y SB3 10 panions of HAeBe stars. Indeed, in hot intermediate-mass T Ori N Y 7.7000 b 100 stars, the Li i absorption line, extremely weak, is not de- HD 53367 N SB1 0.7000 a 30 f tected, whereas in lower mass stars, Li i is detected (see MWC 300 N Y 50 Walter et al. 1988; Duncan & Rebull 1996; Jones et al. AS 442 N SB1 70 00 a,c e 1996). If the spectroscopic signature of this element is MWC 361 N Y 2.50 40 00 a,b,c present in the spectrum of a HAeBe star, it reveals then MWC 1080 N Y 0.80 200 the presence of a young lower mass and cooler companion. a Bouvier et al. 1998; b Leinert et al. 1997b; c Pirzkal et al. 1997; d Bernacca et al. 1993; e B¨ohm & Catala 1995; f Finkenzeller 1985. 2.3.2. Search for radial velocity variations

In order to monitor the radial velocity (Vrad) variations, we mainly used the He i 5 876 and 6 678, Na i 5 890 and 5 895, Si ii 6 347 and 6 371 A˚ lines. The center of the lines was measured by ﬁtting simple Gaussian functions: the errors of such measurements are of the order of 5 to 10 km s−1, depending on the rotational velocity of the stars and its shape (if emission is also present and aﬀects part of the photospheric line).

Note: although with ELODIE´ on-line cross- correlation spectroscopy is possible, we did not used this option: our stars, hot objects with 7 000

i ˚ 3. General results K5). The spectra are shown in Fig. 2 in the Li 6 708 A line region (the rotational velocity for each spectrum is −1 i ˚ 3.1. HAeBe candidates detected as spectroscopic binaries Vrot =50kms ). The Li 6 708 A line, very weak, is not seen in these models (see Gerbaldi et al. 1995; King et al. Table 2 presents the 13 binary systems detected among 1997 for a detailed analysis of the synthetic spectrum of our main sample, i.e. the HAeBe candidates from Table 1 the Li i line in normal in A, F and solar type stars). of the catalogue of Thé et al. (1994). Note: even if such synthetic spectra are only valid for A first group of 6 stars in Table 2 were identified as Main Sequence stars, they help us identifying some in- spectroscopic binaries thanks to the detection of the Li i teresting features in our HAeBe spectra. Indeed, in addi- 6 708 A˚ absorption line. This indicates the presence of a tion to the Li i 6 708 A˚ line, other absorption lines were T Tauri companion. Their spectra are shown in Fig. 1. found to be indicator of a T Tauri companion. Noticeably, Based on stellar models obtained with ATLAS9 we see in Fig. 2 that the absorption feature at 6 678 A˚ (Kurucz 1993) and with solar metallicity, we computed is firstly identified as the Fe i 6 677.989 A˚ line in late- synthetic spectra with SYNSPEC (Hubeny et al. 1994) type stars, and then as the He i 6 678.154 A˚ line when the for various spectral type (B0, B5, A0, A5, F0, F5, G5 and stellar temperature increases towards earlier-type stars. P. Corporon and A.-M. Lagrange: A search for HAeBe spectroscopic binaries 433

Fig. 2. Kurucz models for solar metallicity stars of various spec- Fig. 4. Kurucz models for solar metallicity stars of various spec- −1 −1 tral types (Vrot =50kms )intheLii6 708 Aregion˚ tral types (Vrot =50kms )intheCai6 103 Aregion˚

Fig. 5. Spectra of T Ori in the region of He i 5 875.621, Na i Fig. 3. Ca i lines of HAeBe candidates identified as binaries with 5 889.951 and 5 895.924 A˚ (doted lines show their laboratory positive Li i detection (laboratory positions are shown with position) at three different JDs (Julian Day). Radial velocity doted lines) variations are present in the broad photospheric He i lines and Na i lines,aswellasintheHeiemission feature. Note the strong interstellar Na i absorption lines indicated (IS) The Ca i 6 717.681 A˚ line is seen in late-type stars but not in hotter stars. Thus, the presence in a HAeBe spectrum of the Fe i 6 678 and Ca i 6 718 A˚ lines, even if harder to been recorded. If a radial velocity curve and an orbital detect that the Li i line, are other features that sign the period could be proposed from our observations, the star existence of a cooler companion (see the typical case of is indicated as SB1 for single-lined spectroscopic binary. HK Ori in Fig. 1). Figures 5 to 13 show the spectra of stars with radial ve- Figure 3 shows the Li i 6 103.65 A˚ line region for the locity variations; velocity measurements are gathered in HAeBe stars with positive Li i 6 708 A˚ line detection. No Table 3. For two stars (HD 53367 and AS 442), tenta- evidence of Li i 6 104 A˚ absorption is seen, due to a blend tive radial velocity curves and orbital solutions have been with stronger Fe i and Ca i lines (see also Dunkin et al. obtained using a modified version of the program from 1997). The metallic lines Ca i 6 102.723 and 6 122.217 A˚ are Corporon et al. (1996). absent in spectra of A/B type stars (as shown is the syn- For those stars known to be members of visual binary thetic spectra in Fig. 4), but their presence in our HAeBe systems, we give in the last column of Table 2 their separa- spectra (Fig. 3) are also an evidence for a second lower- tion ρvis. For the visual pairs showing radial velocity vari- mass component. ations, this separation ρvis is probably not related to the The second group in Table 2 is composed of 7 stars for spectroscopic binary separation ρspec <ρvis:athirdcom- which strong evidences of radial velocity variations have ponent is likely involved in those systems, making them 434 P. Corporon and A.-M. Lagrange: A search for HAeBe spectroscopic binaries hierarchical multiple systems. Each star is discussed in Table 3. continued details in Sect. 4. Star JD observed Vrad (−2 400 000) line (km s−1) HD 53367 49769.45 He i 4471 +40.0 3.2. Detected binaries among T2–T5 sample HD 53367 49769.45 He i 5875 +45.8 HD 53367 49769.45 He i 6678 +44.8 Table 4 gives the binary systems spectroscopically de- HD 53367 49821.52 He i 6678 +32.3 tected among HAeBe candidates from Tables 2 to 5 of HD 53367 50087.77 He i 6678 +59.2 the catalogue of Thé et al. (1994). Note that MWC 623 HD 53367 50089.69 He i 6678 +56.3 spectrum in the Li i 6 708 A˚ region, already published in HD 53367 50589.46 He i 6678 +53.4 i Zickgraf & Stahl (1989), is not presented here. HD 53367 50821.67 He 6678 +36.0 i Figures 14 to 16 present the spectra for the spectro- HD 53367 50822.73 He 6678 +35.1 MWC 300 49493.52 He i 5875 −36.3 scopic binaries V361 Ori and HD 199603 and a preliminary MWC 300 49493.52 He i 6678 −33.9 velocity curve for the spectroscopic binary V361 Ori. MWC 300 49493.52 He i 5875 ? +46.4 MWC 300 49493.52 He i 6678 ? +39.5 MWC 300 49591.39 He i 5875 −29.2 3.3. HAeBe candidates undetected as spectroscopic MWC 300 49591.39 He i 6678 −32.9 binaries MWC 300 49591.39 He i 5875 ? +49.9 MWC 300 49591.39 He i 6678 ? +45.3 Table 5 reports all the negative results for Li i line or ra- MWC 300 49875.50 He i 5875 −43.9 dial velocity variations. Some stars with insufficient data MWC 300 49875.50 He i 6678 −38.3 to detect orbital motion are marked with “–” in the cor- MWC 300 49875.50 He i 5875 ? +36.6 ? responding column. MWC 300 49875.50 He i 6678 +27.7 i Figures 17 and 18 show the Li i 6 708 and Ca i 6 103 A˚ MWC 300 49879.47 He 5875 −57.4 i regions for some HAeBe candidates (to be compared with MWC 300 49879.47 He 6678 −47.0 MWC 300 49879.47 He i 5875 ? +37.0 MWC 300 49879.47 He i 6678 ? +34.0 MWC 300 50293.38 He i 5875 39.4 Table 3. HAeBe candidates detected as spectroscopic bina- − MWC 300 50293.38 He i 6678 29.1 ries, through radial velocity variability. Errors are 5 km s−1 for − MWC 300 50293.38 He i 5875 ? +42.0 HD 53367, MWC 300, AS 442, MWC 361, and 10 km s−1 for MWC 300 50293.38 He i 6678 ? +32.6 T Ori, MWC 1080. A symbol “?” indicates a line with emission MWC 300 50633.51 He i 5875 −16.7 Star JD observed Vrad MWC 300 50633.51 He i 6678 −13.7 (−2 400 000) line (km s−1) MWC 300 50633.51 He i 5875 ? +60.5 T Ori 49673.52 He i 5876 +101.6 MWC 300 50633.51 He i 6678 ? +55.2 T Ori 49673.52 He i 5876 ? +21.2 MWC 300 50636.42 He i 5875 −25.0 T Ori 49673.52 Na i 5890 +99.2 MWC 300 50636.42 He i 6678 −17.0 T Ori 49673.52 Na i 5896 +95.4 MWC 300 50636.42 He i 5875 ? +61.1 T Ori 49675.47 He i 5876 +85.1 MWC 300 50636.42 He i 6678 ? +58.4 T Ori 49675.47 He i 5876 ? +18.5 AS 442 49590.53 Mg ii 4481 −30.8 T Ori 49675.47 Na i 5890 +79.9 AS 442 49590.53 Si ii 6347 −28.8 T Ori 49767.35 He i 5876 +87.9 AS 442 49590.53 Si ii 6371 −33.9 T Ori 49767.35 He i 5876 ? +19.0 AS 442 49674.30 Mg ii 4481 −13.2 T Ori 49767.35 Na i 5890 +83.0 AS 442 49674.30 Si ii 6347 −12.7 T Ori 49767.35 Na i 5896 +86.3 AS 442 49674.30 Si ii 6371 −15.6 T Ori 50822.56 He i 5876 +116.4 AS 442 49677.36 Mg ii 4481 −1.5 T Ori 50822.56 He i 5876 ? +75.1 AS 442 49677.36 Si ii 6347 −5.1 T Ori 50822.56 Na i 5890 +71.5 AS 442 49677.36 Si ii 6371 −4.1 T Ori 50822.56 Na i 5896 +70.7 AS 442 49939.41 Mg ii 4481 +3.3 HD 53367 49475.46 He i 6678 +25.1 AS 442 49939.41 Si ii 6347 +9.3 HD 53367 49476.51 He i 6678 +26.6 AS 442 49939.41 Si ii 6371 +6.9 HD 53367 49477.49 He i 5875 +29.2 AS 442 49952.61 Mg ii 4481 +2.6 HD 53367 49647.83 He i 6678 +29.9 AS 442 49952.61 Si ii 6347 +0.8 HD 53367 49648.75 He i 5875 +34.6 AS 442 49952.61 Si ii 6371 +5.0 HD 53367 49673.60 He i 4471 +46.2 AS 442 50637.52 Mg ii 4481 −4.2 HD 53367 49673.60 He i 5875 +49.1 AS 442 50637.52 Si ii 6347 −8.4 HD 53367 49673.60 He i 6678 +47.1 AS 442 50637.52 Si ii 6371 −6.3 HD 53367 49676.72 He i 4471 +46.0 AS 442 50638.49 Mg ii 4481 −1.5 HD 53367 49676.72 He i 5875 +52.3 AS 442 50638.49 Si ii 6347 −7.8 P. Corporon and A.-M. Lagrange: A search for HAeBe spectroscopic binaries 435

Table 3. continued

Star JD observed Vrad (−2 400 000) line (km s−1) AS 442 50638.49 Si ii 6371 −7.0 MWC 361 49591.62 He i 4471 −10.9 MWC 361 49591.62 Mg ii 4481 −21.7 MWC 361 49591.62 He i 5875 −3.3 MWC 361 49591.62 He i 6678 −17.0 MWC 361 49595.45 He i 4471 −12.2 MWC 361 49595.45 Mg ii 4481 −18.3 MWC 361 49595.45 He i 5875 −11.4 MWC 361 49595.45 He i 6678 −17.0 MWC 361 49597.41 He i 4471 −14.0 MWC 361 49597.41 Mg ii 4481 −19.4 MWC 361 49675.36 He i 4471 11.3 − Fig. 6. He i 6 678 A˚ line of HD 53367 at various JDs with radial MWC 361 49675.36 Mg ii 4481 22.9 − velocity variations MWC 361 49675.36 He i 5875 −14.6 MWC 361 49675.36 He i 6678 −17.4 MWC 361 49875.59 He i 4471 −10.7 MWC 361 49875.59 Mg ii 4481 −17.4 MWC 361 49875.59 He i 5875 −8.0 MWC 361 49875.59 He i 6678 −15.7 MWC 361 49940.53 He i 4471 −8.1 MWC 361 49940.53 Mg ii 4481 −25.4 MWC 361 49940.53 He i 5875 −7.9 MWC 361 49940.53 He i 6678 −14.9 MWC 361 49968.41 He i 4471 −7.6 MWC 361 49968.41 Mg ii 4481 −11.0 MWC 361 49968.41 He i 5875 −8.9 MWC 361 49968.41 He i 6678 −15.0 MWC 361 50017.25 He i 6678 −27.0 MWC 361 50293.60 He i 4471 +2.2 MWC 361 50293.60 Mg ii 4481 +2.7 MWC 361 50293.60 He i 5875 +6.6 MWC 361 50635.56 He i 4471 −1.1 MWC 361 50635.56 Mg ii 4481 −5.8 MWC 361 50635.56 He i 5875 −0.3 Fig. 7. Heliocentric radial velocity versus Julian Day for MWC 361 50635.56 He i 6678 −5.2 HD 53367 showing the temporal spread of our observations. MWC 361 50638.50 He i 4471 −0.5 Error bars on the individual points are 5 km s−1 MWC 361 50638.50 Mg ii 4481 −5.4 MWC 361 50638.50 He i 5875 +0.9 MWC 361 50638.50 He i 6678 −3.2 Figs. 1 and 3 which show binary stars). Figure 19 shows MWC 1080 49594.58 He i 5875 −297.8 the He i 6 678 A˚ line of GU CMa with strong variability. MWC 1080 49676.45 He i 5875 −207.7 MWC 1080 49949.60 He i 5875 −190.1 MWC 1080 49953.59 He i 5875 −285.8 3.4. Undetected spectroscopic binaries among T2–T5 sample

Table 6 reports all the negative results for Li i search or radial velocity variations searches for the T2–T5 sample. Table 4. Detected spectroscopic binaries among T2–T5 sample

Star Li i line Vrad Th´e’s vsini detection variations table (km s−1) 3.5. Some stars misclassiﬁed as HAeBe candidates or V361 Ori N SB1 T5 40 doubtful cases MWC 623 Y N T4a HD 199603 N SB1 T5 60 Table 7 presents some stars that, according to the presently available spectra, are very unlikely to be HAeBe candidates. 436 P. Corporon and A.-M. Lagrange: A search for HAeBe spectroscopic binaries

Fig. 10. Si ii 6 347.109 and 6 371.37 A˚ doublet lines in AS 442 Fig. 8. Velocity curve for HD 53367. Error bars on the individ- −1 at two diﬀerent JDs. The division of both spectra ﬁgures the ualpointsare5kms , parameters for the proposed solution radial velocity variation of the doublet lines due to orbital are overplotted motion

Fig. 9. Spectra of MWC 300 in the region of He i 5 875.621, Na i 5 889.951 and 5 895.924 A˚ (doted lines show their labora- Fig. 11. Same as Fig. 8 for AS 442 tory position) at six diﬀerent JDs. Radial velocity variations are present in the broad photospheric He i and Na i lines

4. Notes on individual sources

HK Ori: a cooler companion is detected thanks to the strong Li i 6 708 A˚ line and other metallic lines. This was previously found by Davis et al. (1981). The rotational velocity of the companion is ≈20 ± 3kms−1, the equivalent width (EW) of the Li line is 170 ± 10mA.˚ The broad He i 6 678 A˚ line from the primary is blended with the Fe i 6 680A˚ absorption line from the secondary. No radial velocity variations were recorded during our spectroscopic survey. HK Ori is also known to be a visual binary (separa- 00 tion ρvis =0.34 , see Leinert et al. 1997b). An impor- tantpointisthattheLiidetection and the presence of Fig. 12. Radial velocity variations of the MWC 361 He i 6 678 A˚ Ca i 6 103 and 6 122 A˚ lines of the companion conﬁrms its line P. Corporon and A.-M. Lagrange: A search for HAeBe spectroscopic binaries 437

Fig. 13. Spectra of MWC 1080 in the region of He i 5 875.621 A˚ Fig. 16. Hα 6 563 A˚ line of HD 199603 at various JDs. Variations (doted lines show its laboratory position) at four diﬀerent JDs. in radial velocity are seen. Narrow absorption lines are due to Radial velocity variations are present in the very broad photo- atmospheric H2O spheric He i lines. Strong interstellar Na i absorption lines are indicated (IS)

Fig. 14. Same as Fig. 12 for V361 Ori Fig. 17. HAeBe without Li i 6 708 A˚ line detection

˚ Fig. 15. Same as Fig. 8 for V361 Ori. Star symbols represent Fig. 18. HAeBe (without Li i 6 708 A line) around Ca i 6 103 and ˚ overplotted data from Abt et al. (1991) 6 122 A 438 P. Corporon and A.-M. Lagrange: A search for HAeBe spectroscopic binaries

Table 5. HAeBe candidates undetected as spectroscopic binary. (vsini references in Table 2)

Star Li i line Vrad vsini detection variations (km s−1) HD 245185 N N 150 BF Ori N – 100 e HD 250550 N N 110 e LKHA 215 N – 60 e MWC 147 N N 90 e GU CMa N N ? 150 HEN 3-225 N N 110 f HD 97048 N N 140 e HEN 3-331 N N 50 HEN 3-554 N – HEN 3-644 N – HEN 3-672 N – Fig. 19. He i 6 678 A˚ line of GU CMa at various JDs. Variations HEN 3-692 N – in radial velocity and intensity are observed HD 141569 N N 200 HEN 3-1141 N N 50 Table 7. Stars rejected as HAeBe or doubtful case HR 5999 N – 180 e f Star Li i line Vrad Thé’s vsini MWC 275 N – 120 −1 VV Ser N – detection variations table (km s ) MWC 614 N N 60 RY Ori Y N T2 50 WW Vul N N 125 TCha Y Y? T2 50 HD 190073 N N 20 HD 199603 N Y T5 60 BD+40 4124 N N 180 f MWC 314 N N T4b V361 Cep N N 180 e BD+46 3471 N N 150 e e IL Cep N N 190 youth and lower mass. In the discussion of the system by BH Cep N N 70 Leinert et al. (1997b), the case a) seems appropriate: the SV Cep N N 150 companion is a T Tauri star. A more detailed study of this BHJ 71 N N 25 interesting HAeBe binary system will be given in Bouvier et al. (1998). V380 Ori: as already claimed by Corcoran & Ray Table 6. Undetected spectroscopic binaries in the T2–T5 (1995), Li i line is present in its spectrum: the compan- sample ion has a rotational velocity of 30 ± 5kms−1,EW=80± 10 mAfortheLi˚ iline. Ca i 6 103 and Ca i 6 122 A˚ absorp- Star Li i line Vrad Thé’s vsini detection variations table (km s−1) tion lines from the companion are also detected. V380 Ori 00 HD 45677 N – T3 80 e is a visual binary (ρvis =0.15 ) as well. An analysis of its HD 50138 N – T3 50 characteristics will be presented in Bouvier et al. (1998). HD 97300 N – T5 V586 Ori: while the He i 6 678 A˚ line, due to the HAeBe CD-39 8581 N N T4b a 200 primary, is clearly visible and broad, we also detect for the HD 158352 N N T5 100 first time weak Li i and Ca i 6 717 A˚ lines, indicative of a MWC 925 N – T4b cooler companion. Bouvier et al. (1998) indeed confirm MWC 930 N N T4b 100 the presence of a T Tauri companion at a separation of MWC 953 N N T4b 30 ρ =100. Secondary rotational velocity is about 30 km s−1 AS 321 N N T4b vis and Li i EW is around 50 mA.˚ MWC 314 N N T4b MWC 342 N N T4a NX Pup (A+B): this star is in fact a triple system, 00 MWC 1021 N N T4b 70 consisting of a close binary (ρvis =0.13 ,components MWC 1044 N N T4b 200 A+B, Bernacca et al. 1993; Brandner et al. 1995) asso- 00 MWC 655 N N T4b 200 ciated to a distant companion (ρvis =7 ) (component C). MWC 657 N N T4b 100 We spectroscopically observe the close binary for the fist MWC 1072 N N T4b 200 time. i a This object (= HER 4636) is a visual binary (Chelli et al. While they detected Li line in companion C, Brandner 00 et al. (1995) failed to detect lithium in the A+B pair. 1995) with separation ρvis =4 ; individual components were observed spectroscopically and remarks apply for each star. Our spectrum clearly shows for the first time this line in the binary system, confirming the youth of the system as P. Corporon and A.-M. Lagrange: A search for HAeBe spectroscopic binaries 439 proposed by Schoeller et al. (1996) on the basis of high solution with P ≈ 64 days and e ≈ 0.2, to be confirmed angular resolution optical and near infrared images. Both by other observations. This is a first detection. Li i (FWHM =1.5A,˚ EW = 120 mA)˚ and Ca i lines are MWC 361: some evidences of radial velocity varia- rather broad (FWHM =2A),˚ and may results from a tions are found in various lines such as He i 4 471, 5 876, blend of line from both components. and 6 678 and Mg ii 4 481 A˚ but the data at hand are not Higher S/N spectra with high resolution are needed to enough to set any period. further investigate the spectral type of both components. MWC 1080: a photometric period of P ≈ 2.9days HD 203024:theLiiline has an EW of 70 mA,˚ the has been determined by Shevchenko et al. (1994) and rotational velocity of the low-mass companion (first spec- may compatible with our spectroscopic observations (see troscopic detection) is ≈40 km s−1. HD 203024 is also a spectra at JD = 2449949.60 and JD = 2449953.59 at 00 visual binary with separation ρvis =0.3 (Bouvier et al. nearly opposite phase). Line profile variations complicate 1998). our measurements. Note the very high blue-shift of MWC 863: a cool companion is revealed in Li i as the photospheric lines: Shevchenko et al. (1994) gave a well as in other Ca i line (6 717, 6 103 and 6 122 A).˚ γ velocity of −180 km s−1 for the MWC 1080 binary The Li i has an EW ≈ 40 mA,˚ the rotational velocity system. Together with T Ori, it is probably the third of the secondary is around 30 km s−1. Besides this new eclipsing spectroscopic binary with TY CrA and thus spectroscopic detection, MWC 863 is known to be a deserves further careful observations to constraint the 00 visual binary with ρvis =1.1 (Reipurth & Zinnecker physical parameters of the system. 1993). None of these seven latest stars but TY CrA show a Li i TY CrA: this young triple system has been exten- absorption: the companion must be a low mass star with sively spectroscopically surveyed (Lagrange et al. 1993; low luminosity – typically few percents of the primary lu- Corporon et al. 1994, 1996; Beust et al. 1997) and a re- minosity. A young massif (HAeBe) companion (thus with- cent photometric analysis has been made by Casey et al. out Li i) would have on the other hand a higher luminosity (1998) and Vaz et al. (1998). It consists of a close central and its lines should have been detected, unless highly ob- binary (P ≈ 2.9 days), with a HAeBe primary star and scured by dust. a lower-mass companion; a third much farther away low- Note that HD 53367, MWC 1080 and MWC 361 have mass component orbits the binary. Both lower mass com- all a visual companion. However, these companions are ponents show Li i absorption line (see Casey et al. 1995; unlikely to be responsible for the radial velocity variations Corporon et al. 1996). here observed because of the high separation, but rather TOri: this star has already been reported to be a tertiary component may be involved in each visual a spectroscopic and eclipsing binary by Shevchenko & binary system. If the periodic radial velocity variations Vitrichenko (1994). With the present available data, we of these stars are confirmed, then HD 53367, MWC 1080 observe the radial velocity variations but are unable to and MWC 361 are likely hierarchical multiple systems, as confirm the proposed period P ≈ 14 days for the binary TY CrA. system. Further study of this interesting object is needed to obtain precise masses and radii as it has been done for V361 Ori: emission in the blue part of the He i 6 678 A˚ TY CrA. line is sometimes present, as seen in Fig. 14: this may HD 53367: Herbst & Assousa (1977) and Finkenzeller compromise a good measurement of the central position & Mundt (1984) already reported radial velocity varia- of the line. However, we also made measurements on He i tions for this star. Here, periodic radial velocity variations 4 471 and 5 875, and the more “symmetric” Mg ii 4 481 A˚ are reported for the first time in He i 4 471, 5 876, and absorption lines. Abt et al. (1991) observed this star 6 678 A˚ lines: we propose a period of P ≈ 166 days and an (= Brun 760) but considered it to be constant in radial ve- eccentricity e ≈ 0.18 for the orbital motion of the binary locity. The velocity curve showed in Fig. 15 is computed system. More data are needed to confirm this tentative with our own measurements; some data (those with the orbital solution. lower sigma) from Abt et al. (1991) are overplotted. MWC 300: emission in the red part of He i 5 876 and New measurements are however needed to confirm the 6 678 A˚ makes it difficult to measure of the photospheric spectroscopic binarity status of V361 Ori. central absorption, but radial velocity variations are cor- MWC 623: Zickgraf & Stahl (1989) found this star to related with the Na i 5 890 and 5996 A˚ absorption doublet: be a binary system with a Li-rich K star, but this property, a monitoring of this star is needed to provide an estimate confirmed here, was not underlined in Thé et al. (1994). of the period, our data being to largely spread in time. HD 199603: this star is not a classical HAeBe star, no This is a first detection. emission lines are seen in its spectrum. It is a well-known AS 442: periodic radial velocity variations are found spectroscopic and eclipsing binary of β Lyr type with a pe- in Mg ii 4 481 A˚ doublet as well as in the Si ii 6 347 and riod of P =1.58 days (Pedoussault et al. 1984), a property 6 371 A˚ absorption doublet. We found a possible orbital which was not mentioned in Thé et al. (1994). 440 P. Corporon and A.-M. Lagrange: A search for HAeBe spectroscopic binaries

MWC 147: Vieira & Cunha (1994) classified that this star as a spectroscopic binary with a period P =1year and a circular orbit. They based their study on analysis of Hα spectra, with strong emission: our numerous spectra (15 spectra covering 3 years of observations), do not show any radial velocity variation within the error equal to 5 km s−1, neither in Hα nor in other photospheric absorption lines (He i 4 471, 5 876, Mg ii 4 481 A).˚ Its spectroscopic binarity status is then doubtful. Note that the IR 00 companion (ρvis =3.1 = 900 AU at 290 pc, distance form Hipparcos data) cannot be responsible for the doubtful radial velocity variations observed by Vieira & Cunha (1994), as its orbital velocity would be near 27 103 km s−1 if its orbit is circular. GU CMa: Figure 19 shows the He i 6 678 A˚ line of GU CMa with strongly variable broad features which could sign the possible transit of a rapidly rotating com- Fig. 20. Spectrum of HD 94509, showing a broad He i panion. Other lines He i 4 471, and 5 876 A˚ also show inten- 6 678.154 A˚ line and other narrow metallic lines, probably sity variability, as the Hα line. Adding that Shevchenko originating from the shell. The zoomed windows display the et al. (1992) classified its photometric light curve as quasi- three different narrow lines at JD = 2450821.76, 2450588.48 periodic, this star should be monitored to study possible and 2449822.53 (from top to bottom). Variations in radial velocity are observed circumstellar material transit or chromospheric activity. IL Cep: Shevchenko & Vitrichenko (1994) proposed a photometric period of P = 51 days but found a constant radial velocity within the error equal to 6 km s−1,aswe do. The eclipsing binary status of this star remains to be confirmed. RY Ori: the spectral type is F6 as given by SIMBAD: it may be a composite spectrum binary as pointed out by Herbig & Bell (1988) but we lack good quality spectra in the blue to support this view. The Fig. 21 shows its Li i absorption line: if not due to a companion star, RY Ori is more probably a T Tauri star. TCha: this young PMS star has a G8 spectral type: the Li i line detection confirms its youth (Covino et al. 1997). Possible radial variations may be present, however other data are needed to definitively assert this. HD 94509: this star shows a composite spectrum in ˚ the He i 6 678 A region. Superimposed on the broad He i Fig. 21. Doubtful HAeBe (with Li i 6 708 A˚ line detection) 6 678 A˚ line, narrow absorption features are seen: we ten- tatively identified them as Fe i 6 677.989, Ca i 6 680.628 and Ca i 6 717.681 A˚ lines from a cooler companion, but MWC 314: Miroshnichenko (1996) made a detailed the velocity of each element were not coherent: Fei was analysis of this star and proposed it to be a LBV 1 1 at −45 km s− ,Cai6 681 line at −35 km s− and Ca i candidate: MWC 314 is very unlikely a HAeBe star. 1 6 717 line at −7kms− . Furthermore Li i is not detected, Our data do not suggest neither the presence of radial making doubtful the identification with a young cool com- velocity variations. panion. However, remembering that HD 94509 is an A0I shell star (Reed & Beatty 1995), we may consider that the narrow absorption features are metallic lines formed in the 5. Binary frequency among HAeBe stars shell. If the strong absorption is due to the He i 6 678 A˚ line from the shell, its velocity is around −50 km s−1, while if the two others narrow absorption features are from In the course of this survey, we detected 6 binary sys- Fe i 6 677.989 and 6 715.383 A˚ their respective velocity are tems through Li absorptions and 7 spectroscopic binaries. +90 ± 10 km s−1. Small radial velocity variations may also Before deriving any physical information, we need to ad- be present from one spectrum to another. This star need dress the important question of biases. Several biases have further study to precise its evolutionary status. been identified and are therefore discussed below. P. Corporon and A.-M. Lagrange: A search for HAeBe spectroscopic binaries 441

Fig. 22. The distribution of vsini for the observed HAeBe from Fig. 23. The distribution of V magnitude for the observed Table 1 of Thé et al. (1994). The bin size is 30 km s−1, filled HAeBe from Table 1 of Thé et al. (1994). Filled squared repre- squared represent spectroscopic binaries with radial velocity sent spectroscopic binaries with radial velocity variations from variations from Table 2 Table 2, while open triangles represent spectrum binaries with Li i line detection. Note that TY CrA (V =9.5) shows both radial velocity variations and Li i absorption from companions 5.1. Possible biases

parently) non-binary stars with unknown vsini (probably 5.1.1. Rotational velocity larger than 100 km s−1). Finally, if the T Tauri companion is a fast rotator −1 Herbig Ae/Be stars show usually a broad distribution in (Vrot > 50 − 100 km s ), its Li i line could be more dif- 1 vsini , up to 300 km s− (Grady et al. 1996). The higher ficult to detect. This argument may however be compen- rotational velocity of the HAeBe primary the more diffi- sated by the fact that usually rapid rotators show a larger cult it is to detect radial velocity variations. The projected abundance of Li i than slow rotators do (see Soderblom rotational velocities of our stars are displayed in last col- et al. 1993; Martin et al. 1994; Cunha et al. 1995; Jones umn of Tables 2 and 5 for HAeBe in T1 sample, and in et al. 1996) at least for young G and early-K stars; on Tables 4 and 6 for T2–T5 sample. vsini measurements were the other hand, Duncan & Rebull (1996) found no strong obtained by visually comparing some photospheric lines correlation between Lii and vsini for young stars in Orion. (mainly He i 4 471 and 6 678 A)˚ with synthetic spectra If the system is composed of two HAeBe stars with broadened by rotation (Kurucz 1993; Hubeny et al. 1994). similar luminosities, the blend of the lines (broadened by 1 Error on such estimation should not exceed ± 30 km s− . high rotational velocities) will make difficult their radial For some stars, information was lacking (not enough spec- velocity measurement. tral lines or imprecise spectral type). Our results were fi- In conclusion, we estimate that we may have missed at nally compared with previously published measurements least 50% of the spectroscopic binary with radial velocity (Davis et al. 1983; Finkenzeller 1985; Böhm & Catala variations. 1995; Grady et al. 1996), when available. They appeared to be always consistent with them. Restricting ourself to HAeBe from T1 sample, Fig. 22 5.1.2. Luminosity ratio shows the histogram of the vsini . The projected rotational velocity for the stars ranges from 10 km s−1 (TY CrA) If the primary HAeBe star is much more luminous than its −1 to 200 km s (V380 Ori, HD 141569, MWC 1080). T Tauri companion (∆ mV < 4 − 5), the Li i line (among Interestingly enough, we see that all stars with observed other lines) from the secondary companion is obviously radial velocity variations (symbolized with a filled square) very difficult to detect. So the binary criterion will be the have a vsini below 100 km s−1, but MWC 1080. So, among radial velocity variations of the primary, if any. −1 the 16 stars with vsini< 100 km s , 6 are spectroscopic Figure 23 shows the mV histogram for stars in T1 binaries, i.e. a binary frequency of 37.5%. If we apply sample. HAeBe binary systems with radial velocity vari- this ratio to the 20 remaining high-rotators (100 < ations (filled squares) are rather well distributed between −1 vsini< 200 km s ), we should expect to detect 6to7 mV =6.5 and mV = 11. For spectrum binaries (with Li i more spectroscopic binaries, in addition to MWC 1080. line absorption from a cooler component), there may be a Note that we have not included in this scheme the 5 (ap- lack of detection if mV < 8.5. Keeping the same binary 442 P. Corporon and A.-M. Lagrange: A search for HAeBe spectroscopic binaries frequency as for fainter stars (7 spectrum binaries includ- is 10%. For short-period (P<100 days) WTTS spectro- ing TY CrA among 28 stars with mV > 8.5), up to 4 scopic binaries, Mathieu (1992) found a binary frequency binaries may have been missed among the 14 remaining fb =11±4%, slightly higher than for MS solar-mass brightest stars. stars fb =7±2% found by Duquennoy & Mayor (1991). For spectrum binary systems composed of two HAeBe Our present binary frequency estimate for short period stars, Li i criterion is not anymore valid to probe the du- systems seems comparable to the one for T Tauri or MS plicity. Bouvier and collaborators (Bouvier et al. 1998; stars. However, as our number are small and the biases Corporon 1998) have detected at least 5 pairs of gravita- important, this binary frequency for short-period systems tionally linked HAeBe among the 30 visual binary systems should be regarded as a lower limit: future observations (separation ρ>0.1300) observed with Adaptive Optics. (e.g. using interferometric technics) will certainly help to Assuming a similar ratio for much smaller binary sepa- detect new close systems. rations, we could expect to detect 1 or 2 new spectroscopic binaries composed of two HAeBe stars (this estimate is however a lower limit as not all spectral type for 5.3. Is X-ray emission a binary criterion? visual companions in Bouvier et al. study could have been determined). We address now the puzzling issue of X-ray emission in HAeBe stars. We may wonder if this property could also be used to identify double stars and thus give an other 5.1.3. Branch effect: way of determining the binary frequency. Herbig Ae/Be are known to be strong X-ray emitters This effect has already been described and we showed that (Zinnecker & Preibisch 1994; Damiani et al. 1994). it was not important for binaries with one high mass com- However, the existence of X-ray emission intrinsical to ponent and one low mass companion. The 6 systems de- Herbig Ae/Be stars is still doubtful: these stars indeed tected through Li absorptions belong to this category. For lack convective zones that could create a solar-type the 7 other spectroscopic binary systems, the secondary dynamo and heat a corona via strong magnetic field. spectral type is unknown in most cases. However, the to- A non-solar dynamo model has been proposed by tal luminosity of those systems is much lower than our Tout & Pringle (1995) and applied to the HAeBe star limiting magnitude (mV < 11), except for the faintest HD 104237 by Skinner & Yamauchi (1996): if this shear star MWC 1080: we may have included this spectroscopic dynamo model may generate an active corona, the X-ray binary in our survey because the secondary contributes to luminosity LX predicted seems to be lower than observed. a non-negligible part of the system luminosity. This issue However, many parameters in this model remain free and deserves further studies. are not known empirically. Considering our whole sample, we claim finally that Another possibility is that the X-ray emission detected the Branch effect won’t affect our preliminary binary fre- arises from a cooler T Tauri companion associated to quency estimate for HAeBe stars. the HAeBe star (Zinnecker & Preibisch 1994; Damiani et al. 1994), possibly through a process of colliding winds (Zhekov et al. 1995). In our limited sample of HAeBe bi- 5.1.4. Conclusion on biases: naries (Table 2), 4 stars are known to be X-ray emitters (V380 Ori, TY CrA, MWC 361 and MWC 1080); 2 other In conclusion, at least 50% of the spectroscopic binaries binaries (T Ori and HD 53367) have not been detected could have been missed because of either rotational diffi- by Einstein nor rosat, while the 7 remaining stars have culties measurements or the luminosity ratio. no known X-ray properties. Thus the apparent trend is that X-ray emission is a possible indicator of binarity for HAeBe stars: this conclusion has also been found in 5.2. Derived binary frequency the case of visual HAeBe binaries (Bouvier et al. 1998; Corporon 1998). Nevertheless, it would be worth to ob- Table 2 contains 13 Herbig Ae/Be spectroscopic binaries. serve the 7 remaining binary stars in the X-ray domain. If we only consider the Doppler shift of the lines criterion, we have 7 spectroscopic binary systems (6 stars from the second group of our Table 2 plus the TY CrA system) 6. Conclusion among the 42 HAeBe stars of our principal sample: so our observed binary frequency fb is ≈ 17%.Thisisalower In course of our high resolution spectroscopic survey of limit: for reasons stated above, the true spectroscopic bi- Herbig Ae/Be (HAeBe) binary stars, we observed 42 ob- nary frequency for HAeBe may be as high as 35%. jects with mV < 11: 6 stars exhibit Li i 6 708 A˚ absorp- Restricting ourself to secure or candidate spectro- tion line attributed to a cooler companion (4 are new scopic binary systems with P<100 days (namely T Ori, detections), and for 7 other stars, radial velocity varia- AS 442, MWC 1080 and TY CrA), the binary frequency tions are recorded (4 are new detections). TY CrA is a P. Corporon and A.-M. Lagrange: A search for HAeBe spectroscopic binaries 443 particular triple system for which both binarity criteria Beust H., Corporon P., Siess L., Forestini M., Lagrange A.-M., (i.e. radial velocity variation and presence of Li i absorp- 1997, A&A 320, 478 tion line) are observed. Four stars unlikely appear to be- Böhm T., Catala C., 1995, A&A 301, 155 long to the HAeBe class. Bouvier J., Corporon P., et al., 1998, A&A (in preparation) The Li i line, not observed in our HAeBe candidates Branch D., 1976, ApJ 210, 392 Brandner W., Alcala J.M., Kunkel M., Moneti A., Zinnecker from Table 5, is not only an indicator of binarity, but may H., 1996, A&A 307, 121 also help in some cases to classify a star as either a T Tauri Brandner W., Bouvier J., Grebel E.K., Tessier E., De Winter or as a HAeBe star. Depending on the presence or not of D., Beuzit J.L., 1995, A&A 298, 818 the Li i 6 708 A˚ line, the distinction, between the class of Casey B.W., Mathieu R.D., Suntzeff N.B., Lee C.-W., Cardelli young low mass stars and the class of young intermedi- J.A., 1993, AJ 105, 2276 ate mass stars, could be inferred thanks to this criterion. Casey B.W., Mathieu R.D., Suntzeff N.B., Walter F.M., 1995, This distinction has been proposed here for some peculiar AJ 109, 2156 cases, such as RY Ori or T Cha. Casey B.W., Mathieu R.D., Vaz L.P.R., Andersen J., Suntzeff Within our reduced sample, the observed binary N.B., 1998, AJ 115, 1617 Chelli A., Cruz-Gonzalez I., Reipurth B., 1995, A&AS 114, frequency for short-period spectroscopic HAeBe systems 135 (fb ≈ 10%) is roughly comparable to the one of T Tauri Clarke C., 1996, “Evolutionnary processes in binary stars” or MS stars. The true binary frequency is probably higher Wijers A.M.J., Melvyn B.D. and Tout C.A. (ed.), NATO considering to the present biases working against the ASI Series, Vol. C 477, p. 31 detection of spectroscopic binary stars. Corcoran D., Ray T.P., 1995, A&A 301, 729 Due the limitation in magnitude, this systematic Corporon P., 1998, Ph.D. thesis, Université Grenoble I search would be greatly completed with the use of 8 m- Corporon P., Lagrange A.-M., Beust H., 1996, A&A 310, 228 class telescope, to have access of fainter HAeBe stars and Corporon P., Lagrange A.-M., Bouvier J., 1994, A&A 282, L21 enlarge our sample, and to interferometric technics under Covino E., Alcala J.M., Allain S., Bouvier J., Terranegra L., Krautter J., 1997, A&A 328, 187 development. It is emphasized that the spectroscopic Cunha K., Smith V.V., Lambert D.L., 1995, ApJ 452, 634 binaries discussed in the paper deserves further obser- Damiani F., Micela G., Sciortino S., Harnden F.R.J., 1994, vations in order to obtain more information about the ApJ 436, 807 secondary component (luminosity, age) and possibly to Davis R., Strom K.M., Strom S.E., 1983, AJ 88, 1644 retrieve the masses: their careful study by various means Davis R.E., Strom S.E., Strom K.M., 1981, BAAS 13, 855 (spectroscopic follow-up, careful photometric monitoring, Duncan D.K., Rebull L.M., 1996, PASP 108, 738 lunar occultation, interferometric measurements...) are Dunkin S.K., Barlow M.J., Ryan S.G., 1997, MNRAS 290, 165 encouraged. Duquennoy A., Mayor M., 1991, A&A 248, 524 Figueiredo J., 1997, A&A 318, 783 Finkenzeller U., 1985, A&A 151, 340 Acknowledgements. This research has made use of the Simbad Finkenzeller U., Mundt R., 1984, ApJS 55, 109 database, operated at CDS, Strasbourg, France. We warmly Garmany C.D., Conti P.S., Massey P., 1980, ApJ 242, 1063 thanks the INSU and ESO observing programs committees for Gerbaldi M., Faraggiana R., Castelli F., 1995, A&AS 111, 1 generous telescope time allocations and the OHP and ESO tele- Ghez A.M., Mc Carthy D.W., Patience J.L., Beck T.L., 1997, scope teams for their kind support. P.C. thanks P. Prugniel ApJ 481, 378 for early discussion at OHP and D. Queloz for his efficient Ghez A.M., Neugebauer G., Matthews K., 1993, AJ 106, 2005 and friendly presence during the first nights with ELODIE.´ Gillet D., Burnage R., Kohler D., et al., 1994, Ap&SS 108, 181 We would like to express our gratitude towards S. Allain, Grady C.A., Pérez M.R., Talavera A., et al., 1996, A&AS 120, D. Mouillet, J.C. Augereau and L. Siess who kindly per- 157 formed some of the observations, and also to C. Ounnas and J. Herbig G.H., Bell K.R., 1988, Lick Obs. Bulle. 1111, 1 Rodriguez for some service observations. We thank J. Bouvier Herbst W., Assousa G.E., 1977, ApJ 217, 473 for providing useful comments, and H. Beust who specially Hillenbrand L.A., 1994, in “The nature and evolutionary sta- modified his program for this study. Finally, we are grateful to tus of Herbig Ae/Be stars”, Thé P.S., Pérez M.R., van den the referee, Dr. Andrea Richichi, for useful comments on this Heuvel E.P.J (eds.), ASP Conf. Ser. 62, 369 work and careful reading of the manuscript. Hubeny I., Lanz T., Jeffery C.S., 1994, News1. Anal. Astron. Spectra 20, 30 Jones B.F., Shetrone M., Fischer D., Soderblom D.R., 1996, AJ 112, 186 References King J.R., Deliyannis C.P., Hiltgen D.D., Stephens A., Cunha K., Boesgaard A.M., 1997, AJ 113, 1871 Abt H.A., Gomez A.E., Levy S.G., 1990, ApJS 74, 551 Kurucz R.L., 1993, Cambridge, SAO Abt H.A., Wang R., Cardona O., 1991, ApJ 367, 155 Lada E.A., Evans N.J., Depoy D.L., Gatley I., 1991, ApJ 371, Baranne A., Queloz D., Mayor M., et al., 1996, A&AS 119, 171 390 Lagrange A.-M., Corporon P., Bouvier J., 1993, A&A 274, 785 Bernacca P.L., Lattanzi M.G., Bucciarelli B., et al., 1993, A&A Leinert C., Henry T., Glindemann A., Mc Carthy D.W., 1997a, 278, L47 A&A 325, 159 444 P. Corporon and A.-M. Lagrange: A search for HAeBe spectroscopic binaries

Leinert C., Richichi A., Haas M., 1997b, A&A 318, 472 421, 517 Leinert C., Zinnecker H., Weitzel N., Christou J., Ridgway Reed B.C., Beatty A.E., 1995, ApJS 97, 189 S.T., Jameson R., Haas M., Lenzen R., 1993, A&A 278, Reipurth B., Zinnecker H., 1993, A&A 278, 81 129 Schoeller M., Brandner W., Lehmann T., Weigelt G., Zinnecker Li W., Evans N.J., Harvey P.M., Colome C., 1994, ApJ 433, H., 1996, A&A 315, 445 199 Shevchenko V.S.N., Grankin K.A., Ibragimov M.B., Martin E.L., 1994, in “The nature and evolutionary status Kondratiev V.Yu., Melnikov S., 1994, in “The nature of Herbig Ae/Be stars”, Thé P.S., Pérez M.R. et van den and evolutionary status of Herbig Ae/Be stars” Thé P.S., Heuvel E.P.J. (eds.), Vol. 62, p. 315 Pérez M.R. et van den Heuvel E.P.J. (eds.), Vol. 62, p. 43 Martin E.L., Rebolo R., Magazzu A., Pavlenko Y.V., 1994, Shevchenko V.S., Grankin N., Ibragimov K.A., Yu. M., A&A 282, 503 Melnikov S., Yakubov D., 1992, Astrophys. Space Sci. 202, Mathieu R.D., 1992, in “A Complementary Approaches to 121 Double and Multiple Star Research”, McAlister H.A., Shevchenko V.S., Vitrichenko E.A., 1994, in “The nature and Hartkopf W.I. (eds.), IAU Colloquium 135, ASP Conf. evolutionary status of Herbig Ae/Be stars” Thé P.S., Pérez Ser. 32, 30 M.R., van den Heuvel E.P.J. (eds.), Vol. 62, p. 55 Mayor M., Duquennoy A., Halbwachs J.L., Mermilliod J.-C., Skinner S.L., Yamauchi S., 1996, ApJ 471, 987 1992, in “A Complementary Approaches to Double and Simon M., Ghez A.M., Leinert C., et al., 1995, ApJ 443, 625 Multiple Star Research”, McAlister H.A., Hartkopf W.I. Soderblom D.R., Jones B.F., Balachandran S., Stauffer J.R., (eds.), IAU Colloquium 135, ASP Conf. Ser. 32, 73 Duncan D.K., Fedele S.B., Hudon J.D., 1993, AJ 106, 1059 Miller G.E., Scalo J.M., 1978, PASP 90, 506 Thé P.S., de Winter D., Perez M.R., 1994, A&AS 104, 315 Miroshnichenko A.S., 1996, A&A 312, 941 Tout C.A., Pringle J.E., 1995, MNRAS 272, 528 Morse J.A., Mathieu R.D., Levine S.E., 1991, AJ 101, 1495 Vaz L.P.R., Andersen J., Casey B.W., Clausen J.V., Mathieu Nordström B., Stefanik R.P., Latham D.W., Andersen J., 1997, R.D., Heyer I., 1998, A&AS 130, 245 A&AS 126, 21 Vieira S.L.A., Cunha N.C.S., 1994, IAU Informational Bulletin Padgett D.L., Stapelfeldt K.R., 1994, AJ 107, 720 of Variable Stars 4090, 1 Padgett D.L., Strom S.E., Ghez A., 1997, ApJ 477, 705 Walter F.M., Brown A., Mathieu R.D., Myers P.C., Vrba F.J., Pedoussaut A., Ginestet N., Carquillat J.M., 1984, A&AS 58, 1988, AJ 96, 297 601 Welty A.D., 1995, AJ 110, 776 Pirzkal N., Spillar E.J., Dyck H.M., 1997, ApJ 481, 392 Zhekov S.A., Palla F., Prusti T., 1995, MNRAS 276, L51 Prosser C.F., Stauffer J.R., Hartmann L., Soderblom D.R., Zickgraf F.J., Stahl O., 1989, A&A 223, 165 Jones B.F., Werner M.W., Mc Caughrean M.J., 1994, ApJ Zinnecker H., Preibisch T., 1994, A&A 292, 152